Short synthesis of 2,3,4,5-tetrahydrocytisine

2,3,4,5-Tetrahydrocytisine ( 5 ) was synthesized from quinolizidine-2,6-dione in three steps and acceptable yield.


Introduction
Due to its ability to affect the nicotinic cholinergic receptors 1 selectively, cytisine 1, a well known and widespread representative of quinolizidine alkaloids 2 of family Leguminosae has attracted considerable attention in the pharmacology 3,4 and in the synthetic chemistry, as well.Especially, the recent results 5,6,7,8 of pharmacological studies stimulated the synthetic research considerably, inspiring the development of new drugs, e.g. in smoking cessation.As a result of the numerous trials some different strategies have been elaborated towards its total synthesis starting mainly from properly substituted pyridines 9,10,11 , piperidines 12 or bispidines 13 , affording the target molecule at the end of multistep procedures in poor or moderate yields.These methods have been surveyed in a comparative manner in an excellent review published recently. 14O NH Cytisine (1)   Some saturated compounds including the 3,7-diazatricyclic moiety of cytisine were synthesized from 2-quinolizidone in our laboratory some years ago.15 Our experiments to convert these saturated diazatricyclics to cytisine proved, however, unsuccessful, though a wide scale of oxidizing agents, e.g.potassium hexacyanoferrate(III), sodium dichromate, mercury(II) acetate, 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DCCQ), was tried to effect the regioselective tertiary amine → lactam oxidation.Unfortunately, no satisfactory selectivity was observed, and the very complex mixtures obtained in these oxidative transformations made this strategy useless.

Results and Discussion
Due to the difficulties mentioned above quinolizidine-2,6-dione (2), a starting compound including the lactam functionality, was subjected to a double Mannich condensation used successfully in our previous work. 15he key intermediate 2 was synthesized using the biomimetic α-acylimmonium cyclization method. 16Thus, glutarimide was alkylated with 3-butyn-1-ol in a Mitsunobu protocol and reduced with L-Selectride R , followed by treatment with formic acid, to give the desired product in an acceptable yield.An alternative synthetic route was also used to obtain 2. The 2-piperidone diester available from simple chemicals in some steps 17 was cyclized, hydrolyzed and decarboxylated in a usual Dieckmann sequence to produce 2 on a multigram scale without need of the chromatographic purification (Scheme 1).The double Mannich condensation of 2 with formaldehyde and methylamine furnished the ketones 3a,b in good yield (Scheme 2).The reduction of the C(13) carbonyl group of 3b was effected through reduction of its tosylhydrazone with sodium cyanoborohydride in sulfolandimethyl formamide mixture to yield 4. The classical Wolff-Kishner procedure gave the same product without any observable damage of the lactam group and in a higher yield (Scheme 3).Removal of the benzyl group of 4 in the usual way (hydrogenation with palladium/charcoal) resulted in 2,3,4,5-tetrahydrocytisine (5) as a racemic mixture.Reduction of the ketolactam 3b with sodium borohydride led to the formation of an isomeric pair of hydroxylactams 6 (α isomer) and 7 (β isomer) in a ratio 2/8.Based on a previous NMR signal assignment 15,18 the steric position of the C(13)-OH group was determined by 13 C NMR spectroscopy.

Conclusions
In this work we demonstrated a new synthetic strategy to construct the diazatricyclic skeleton of cytisine and related compounds.The easy access of the intermediate 2 and the simple transformations provide a short and efficient pathway in total synthesis of cytisine and its derivatives.

Experimental Section
General Procedures.All solvents purchased commercially were redistilled and dried before use.TLC was made with Kieselgel 60F 254 Plastikfolien from Merck.Merck Kieselgel 60 (0.040-0.063 mm) was used for column chromatography.Melting points were measured on a Büchi SMP 20 apparatus and not corrected.IR spectra: Perkin-Elmer FTIR 1600 spectrometer, 1 H and 13 C NMR spectra: Bruker AM-300 at 300 and 75 MHz, or Varian VXR at 400 and 100 MHz, respectively.
Trituration with diethyl ether gave 3,81 g of 3b as a slightly yellow solid.Flash chromatography of the residue (eluent ethyl acetate -acetone 1:1) furnished an additional portion of 1,68 g of 3b.